CARBON DIOXIDE EMISSIONS, ENERGY CONSUMPTION, AND ECONOMIC GROWTH IN A TRANSITION ECONOMY: EMPIRICAL EVIDENCE FROM CAMBODIA

CARBON DIOXIDE EMISSIONS, ENERGY CONSUMPTION, AND ECONOMIC GROWTH IN A TRANSITION ECONOMY:

EMPIRICAL EVIDENCE FROM CAMBODIA

Tuck Cheong Tang^a1, Pei Pei Tan^a

aDepartment of Economics, Faculty of Economics & Administration, University of Malaya

Abstract

This study examines the inter-relationship among carbon dioxide (CO2) emissions, energy consumption, and economic growth for a Mekong River Commission (MRC) country - Cambodia. The empirical results suggest that real gross domestic product (GDP), energy consumption, and CO2 emissions are cointegrated. It needs 11 years to achieve a long-run equilibrium. There is a unidirectional causality from real GDP to energy consumption, and a bidirectional causality between real GDP, and CO2 emissions. The CO2 emissions are related to energy consumption through real GDP. This study is relevant and importance for Cambodia in formulating energy policies, for example, the revision of national energy efficiency policy.

JEL Classification: C22, Q43, Q48.

Keywords: Cambodia; Causality; Cointegration; CO2 emissions; Energy consumption; Economic growth.

1. Introduction

The research issue inspires this study is the current phenomenon reported by the Phnom Penh Post that “Cambodia is suffering disproportionately from the impacts of greenhouse gas emissions from more developed nations, according to a new study published in the journal Nature on Friday… Cambodia is one of 36 countries “severely” affected by global climate change, as of 2010. If current trends continue, Cambodia’s vulnerability will downgrade slightly to “acute” by 2030” (The Phnom Penh Post, 8 February 2016).² The Cambodia's power consumption is forecasted to rise to 3.4 TWh by the end of 2020 to achieve at       

1 Corresponding author: E-mail: tangtuckcheong@um.edu.my.

2 http://m.phnompenhpost.com/national/polluters-hurt-kingdom-

study?utm_content=bufferd30f0&utm_medium=social&utm_source=facebook.com&utm_cam paign=buffer (Accessed: March 1, 2016)

9.4% growth (Royal Government of Cambodia, 2013, p. 2). Meanwhile, the Royal Government of Cambodia (2013) has highlighted that the Cambodian annual electricity demand has increased about 16.3% from 2002 to 2011 and, the CO2

emissions from energy consumption has amounted to nearly 4 million tonnes.

More precisely, the primary energy consumption and CO2 emission at least doubled over the past ten years – it is eventually a major challenge for the national energy policy. In fact, the overall policy goal of the Cambodian energy efficiency is to reduce the future national demand for energy by 20% at 2035, as well as national CO2 emissions in 2035 by 3 million tonnes of CO2 (Royal Government of Cambodia, 2013, p. 7).

Figure 1 (the top plot) shows that the Cambodian CO2 emissions and real GDP increase substantially since 1995. They are closely correlated - higher real GDP causes more CO2. The Cambodian real GDP ‘takes off’ in 1985 as a result of economic reforms since the past two decades from a command economy in the late 1980s to a free market economy in the recent (Tang and Chea, 2013).

According to the United Nations (2003), Cambodia is one of the most open economies in the Southeast Asia region³ and has been labelled as one of the new tiger economies of Asia, according to the forecast in the Asian Development Bank’s Asian Development Outlook 2016.⁴ The bottom plot shows the oil consumption increases since 1994 but drops drastically in 2008. It is consistent with the second structural shift of CO2 emissions in 1995 (the first was in 1983) suggesting a positive correlation. Visual inspection of the plot shows that oil and electricity consumptions, CO2 emissions, and real GDP are positively associated.

The electricity used gradually increases since 2001, while the primary energy consumption is relative stable. The International Energy Statistics 2012 reports that in 2009 the CO2 emissions from energy consumption amounted to 3.93 million tonnes that both figures the demand for primary energy, and CO2

emissions at least doubled over the past ten years (Royal Government of Cambodia, 2013, p. 2). Hence, a case study of a Mekong River bordering country, Cambodia is gaining considerable interest among the researchers on the relationships among energy consumption, CO2 emissions and economic growth.

The present issue in Cambodia is related to other countries from the previous studies, especially for the regional energy study of Association of Southeast Asian Nations (ASEAN). Cambodia joined the ASEAN on 30 April 1999. Cambodia plays a virtual role in terms of intra-regional co-operation that the ASEAN countries have an active agenda on many energy policy fronts, and they are together continuously to strive towards implementation of long-standing projects in order to establish interconnected grids for electricity and natural gas (namely       

3 It is based on the economic freedom index compiled by the Heritage Foundation in the United States. Cambodia is ranked 35^th among 170 countries for the year of 2003. The rankings for its neighbouring countries are 72^nd for Malaysia, 99^th for Indonesia, 135^th for Vietnam, and 153^rd for Lao People’s Democratic Republic. Cambodia is ranked among the world’s least developed countries (LDCs) at the very top in market-friendliness. Cambodia has offered a set liberal policies to investors. (United Nations, 2003).

4 http://www.adb.org/news/features/here-comes-cambodia-asia-s-new-tiger-economy (Accessed: September 5, 2016)

the ASEAN Power Grid and the Trans-ASEAN Gas Pipeline) - but each country has its own key policies and targets.⁵ Therefore, panel approach on the energy study with Cambodia such as Lee and Brahmasrene (2014) is infeasible. In fact, no study is available for a case study of Cambodia. This study contributes to the empirical literature on the empirical evidence of energy-CO2-growth nexus for Cambodia.

Figure 1: Plots of real GDP, CO2 emissions, and energy consumption from 1980-2010.

      

5 See “Key energy policies, targets and objectives in ASEAN”, in Table 1.5 (pages 32-33), Southeast Asia Energy Outlook – World Energy Outlook Special Report.

https://www.iea.org/publications/freepublications/publication/SoutheastAsiaEnergyOutlook_

WEO2013SpecialReport.pdf (Accessed: September 5, 2016).

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

Co2 emissions metric tons of Co2 (m)

Real GDP lc (m)

0 0.5 1 1.5 2 2.5

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 0 5 10 15 20 25 30 35 40 Electricity net consumption

kilwatthours (b)

Primary energy consumption guadrillion btu

Oil consumption 000 barrels per day

Only Lee and Brahmasrene (2014)’s study considers Cambodia for a panel data (1991-2009) of 9 ASEAN countries, and they find that information communications technology (ICT), CO2 emissions and economic growth are cointegrated. ICT has a significant positive impact on both economic growth and CO2 emissions. Economic growth and CO2 emissions have feedback causation. A recent study, Wang et al. (2016a) examine the effects of urbanisation on energy consumption, and carbon emission in the 8 ASEAN member countries, namely Singapore, Malaysia, Indonesia, Thailand, the Philippines, Brunei, Vietnam, and Myanmar. The panel cointegration tests suggest long-run relationship for 1980- 2009. A 1% increases in urban population results in a 0.20% higher carbon emission. Urbanisation with energy use causes carbon emission in the long-run.

In the short-run, urbanisation causes both energy use and carbon emission. Baek (2016) investigates the impact of inward foreign direct investment (FDI) on the CO2, GDP, and energy consumption for 5 ASEAN countries (Indonesia, Malaysia, the Philippines, Singapore, and Thailand) for the period of 1981-2010. The inflow of FDI increases CO2 emission. Income and energy consumption have a negative impact on reducing CO2. Heidari et al. (2015) support environmental Kuznets curve (EKC) for a panel of 5 ASEAN (Indonesia, Malaysia, the Philippines, Singapore, and Thailand) over the period of 1980-2008. The panel smooth transition regression (PSTR) shows energy consumption increases CO2 if the GDP per capita is below 4686USD. Chandran and Tang (2013b) find cointegration between CO2 emissions and other variables for Indonesia, Malaysia and Thailand for the period 1971-2008, and economic growth plays a greater role in CO2. Inverted U-shape EKC is not supported in the case of Indonesia, Malaysia and Thailand.

Next section is the literature review with an update of 56 empirical studies between 2015 and 2016. Section 3 describes the data and their degree of integration. Time series testing methods are included in this section – autoregressive distributed lag (ARDL) approach. Section 4 reports the empirical results. Section 5 concludes the study.

2. Literature review - an update

Generally, two hypotheses are being tested empirically in the past studies, namely energy-growth nexus, and CO2-energy-growth nexus. The most common Cobb–

Douglas production function is being applied on the influence of energy consumption on output, while the EKC relates the pollution to output, and the Grey theory proposes a relationship between energy consumption and pollution.

The other studies utilise the consumption theory which relates income and

‘energy’ variable(s) to consumption of goods and services. A seminal work by Kraft and Kraft (1978) documents that Gross National Product (GNP) does cause energy for the postwar period. Most of the past studies are summarised and reported by Ozturk (2010), Mohammadi and Parvaresh (2014), Chandran and Tang (2013a) and (2013b). They conclude that different findings when different sample countries, methods of analysis, and additional variables being considered.

This study summarises a total of 57 articles available between 2015 (29 articles) and 2016 (28 articles) (see Appendix A).⁶

The updated literature review gives several similarities are observed. Firstly, the studies are mainly to (re-)examine the cointegration and causality between economic growth, energy consumption, and pollutants (CO2 emissions). 22 out of 28 articles published in 2016 test for cointegration and causality, except for Bae et al. (2016), Baek (2016), Fujii and Managi (2016), Kais and Sami (2016), Sumabat et al. (2016) and Wang et al. (2016b) on its effects of the variables.

Secondly, they employ a multivariate framework than of bivariate framework by adding new variables such as energy prices, financial development, FDI, health quality, urbanisation, trade openness (international trade), tourism receipts, and so on. All of the studies in 2016 and 21 out of 29 studies in 2015 have considered additional variables, except for Apergis (2016). Thirdly, the single country study is still of the interest in energy study that almost half (28 articles) of the latest studies. The case study is, for example, Pakistan, Greek, Italy, Malaysia, and so on. Finally, the ARDL approach is a widely applied method for testing the cointegration.

3. Data, degree of integration and methods

This section describes the data, their degree of integration, I(d), and testing methods. The four variables are real GDP (Y, in local currency, million), primary energy consumption (PEC, in btu), oil consumption (OC, in ’000 barrels per year), electricity net consumption (ENC, in kilowatt-hours), and CO2 emissions (metric tonnes). Real GDP data are from Tang and Chea (2013), while the energy data are taken from the U.S. Energy Information Administration (http://www.eia.gov/). The sample period is between 1980 and 2010 (annual data).⁷ All of the variables are transformed into natural logarithm (ln).

Table 1 reports the augmented Dickey-Fuller (ADF) (Dickey and Fuller, 1979), Phillips-Perron (PP) (Phillips and Perron, 1988), and Kwiatkowski–Phillips–

Schmidt–Shin (KPSS) (Kwiatkowski et al., 1992) tests. Both of the ADF and PP tests consistently fail to reject the null hypothesis of a unit root for all candidate variables, but they reject the null hypothesis in the first-differenced transformation.⁸ However, a more powerful test, KPSS method rejects the null hypothesis of stationary for all of the variables in levels, but none in first- differences suggesting I(1), except for lnENC which is I(2) i.e. the null of the first- differenced stationary is rejected. Also, in panel (b), the KPSS tests show lnPEC, lnOC and lnCO2 are stationary in levels, or I(0) since the test statistics fail to reject null of trend stationary.

      

6 Other studies include Tang (2008; 2009), Tang and Tan (2012; 2013), Tang and Muhammad (2013), Tang and Tan (2014), and so on.

7 The energy data for Cambodia are available until 2010. The Cubic interpolation has initially considered which generated 121 observations for the periods 1980Q4 – 2010Q4. However, a reservation is the underlying series are not smoothly trended resulting bias in interpolated series.

8 If all variables are I(1) as suggested, vector error correction model (VECM) can be used for short- run as well as long-run model, including causality tests.

Table 1: Unit root and stationary tests.

ADF PP KPSS Level 1st Difference Level 1st Difference Level 1st Difference

Panel (a) No Trend

lnY -2.230 -2.700^* -1.730 -2.691^* 0.645^** 0.297

lnPEC -0.550 -5.032^*** -0.556 -5.032^*** 0.690^** 0.112

lnOC -0.773 -4.856^*** -0.785 -4.856^*** 0.684^** 0.128

lnENC 2.996 -5.819^*** 4.822 -5.237^*** 0.698^** 0.636^**

lnCO2 -2.185 -5.481^*** -2.412 -5.481^*** 0.710^** 0.207 Panel (b) With Trend

lnY 0.058 -3.352^* -0.525 -3.248^* 0.176^** 0.104

lnPEC -1.761 -4.930^*** -1.900 -4.930^*** 0.100 0.112

lnOC -1.561 -4.771^*** -1.735 -4.771^*** 0.096 0.122

lnENC -1.087 -7.278^*** -0.742 -19.249^*** 0.197^** 0.500^***

lnCO2 -2.732 -5.805^*** -2.505 -5.838^*** 0.114 0.082 Notes: For ADF test, Schwarz information criterion (SIC) is used to select the lag length. For PP test, the Barlett kernel is used for the spectral estimation method by using the Newey-West bandwidth. (^***), (^**) and (^*) indicate 1%, 5% and 10% significance level, respectively. In panel (b), the italic statistics have non-significant trend component. The critical value of the finite sample KPSS critical values are obtained from Table 1 and Table 2 from Hornok and Larsson (2000).

This observation dissatisfies the application of the conventional cointegration such as Johansen multivariate cointegration method (Johansen and Juselius, 1990) which requires all underlying variables be I(1). The existence of I(1) variables allows the ARDL approach (Pesaran et al., 2001). The ARDL bounds test is applicable irrespective of whether the independent variables are stationary, I(0) or non-stationary, I(1). It avoids the pre-testing problems associated with conventional cointegration methods that require the degree of integration of the underlying variables either I(1) or I(0), see Pesaran and Pesaran (1997, pp. 302- 3). The lnENC variable is dropped from cointegration analysis since no cointegration among real GDP, ENC, and CO2 emissions can be concluded.⁹

This study follows the empirical framework as employed by Tang et al. (2013), Tang and Salah (2014), Abdul (2014), Wolde-Rufael (2014), and Ruhul et al.

      

9 This procedure (ARDL) is applicable at most I(1) variables. Haldrup (1998) surveys the recent literature dealing with I(2) variables. Standard remedy of differencing the I(2) variable twice, may result in information loss. The Engle-Granger tests cannot reject the null hypothesis of the series are not cointegrated with a maximum lag of 3 (see the statistics below).

Dependent variable tau-statistic Prob. z-statistic Prob.

ENC -0.661 0.985 -1.680 0.986 RGDP -2.276 0.614 -13.240 0.300 CO2 -2.942 0.305 -11.187 0.443

(2014). The relations are lnY -lnOC -lnCO2, lnY –lnPEC -lnCO2, and lnY –lnENC -lnCO2. Following the ARDL modelling framework, a relation of lnY -lnOC - lnCO2, for example, can be written as equation (1).

Δ ∑ Δ ∑ γ Δ ∑ δ Δ

(1)

The computed F-statistic is a restriction of the estimated coefficients of the level variables, lnYt-1, lnOCt-1 and lnCO2t-1 to zero, or to test the null hypothesis of

: 0 (i.e. no long-run relationship between the underlying

variables). This test statistic has a non-standard distribution irrespective of whether lnY, lnOC, and lnCO2 are I(0) or I(1). If the F-statistic falls outside the upper-bound critical values, the null hypothesis can be rejected. It suggests a long-run relation. No long-run relation can be concluded given that the F-statistic is below the lower-bound critical values. Inconclusive inference is delivered, if the F-statistic falls between the lower-and upper-bound critical values and it depends on whether the underlying variables are I(0) or I(1) (Pesaran and Pesaran, 1997, p. 304). The remaining ARDL specifications with energy consumption lnOC can be re-arranged as equations (2) and (3). Similar testing procedure (as equation 1) is applied.

Δ ∑ Δ ∑ γ Δ ∑ δ Δ

(2)

Δ ∑ Δ ∑ γ Δ ∑ δ Δ

(3)

Once a cointegration is suggested, an error correction model (ECM) can be estimated by ordinary least squares (OLS) estimator (Engle & Granger, 1987). It is a restricted version of ARDL equations (1)-(3) in which an error correction term, ecmt-1 replaces the lagged one level variables, where ecmt-1 =

as in equation (1). The quations (1)-(3) are rewritten in ECM form, i.e. equations (1’), (2’) and (3’).

Δ ∑ Δ ∑ γ Δ ∑ δ Δ (1’)

Δ ∑ Δ ∑ γ Δ ∑ δ Δ

(2’)

Δ ∑ Δ ∑ γ Δ ∑ δ Δ

(3’)

On the other hand, Granger non-causality test (Granger, 1988) is employed in order to identify the directions of causality between the variables. According to Engle and Granger (1987, p. 251), “An individual economic variable, viewed as a time series, can wander extensively and yet some pairs of series may be expected to move so that they do not drift too far apart. Typically economic theory will propose forces which tend to keep such series together”. Toda and Yamamoto (1995) method is used in this study because it allows non-causality test without pre-testing cointegration either the underlying variables are cointegrated or non-cointegrated of an arbitrary order. Also, it permits a mixture integration of the variables whether a series is I(0), I(1) or I(2) such as the case of this study that lnENC is I(2). The details of this widely applied method are available from Toda and Yamamoto (1995). It involves two steps; (i) determine the true lag length of k and the maximum order of integration (dmax) of the underlying variables in the system, and an augmented VAR(k + dmax) is then estimated by OLS estimator; and (ii) compute the standard Wald tests to the first kth VAR coefficient matrix only or to test for restrictions on the parameters of the VAR(k) model in order to reject the null hypothesis of ‘x does not Granger-cause y’. The test statistic follows an asymptotic chi-squared distribution with k degrees of freedom in the limit when a VAR(k + dmax) is estimated.

4. Empirical results

This section reports the empirical results. Table 2 presents the computed F- statistics of the ARDL specifications (a) - (f) for cointegration and their critical values that consider a small sample of 30 observations. As noted in the previous section, lnENC is I(2), and no cointegration. The F-statistics of all specifications, except for CO2 equations (c) and (f) exceed the upper bound of the critical value, 3.695 at 10% significance level. Hence, the null hypothesis of ‘there exists no long- run relationship between real GDP, energy consumption, and CO2 emissions’ can be rejected, irrespective of the order of their integration I(0) or I(1). It suggests that energy consumption, real GDP, and CO2 emissions are cointegrated with the following relations, i.e. lnY -lnOC -lnCO2, lnY -lnPEC -lnCO2, lnOC -lnY -lnCO2, and lnPEC -lnY -lnCO2.¹⁰

The estimated relations (a) – (f) are reported in Table 3 by the ARDL approach. They pass a set of diagnostic checking for serial correlation, function form, normality, and heteroscedasticity. The two key relations (a) and (d) show both energy consumption and CO2 emissions are statistically insignificant to explain real GDP in the long-run. An increase in oil consumption (0.970) and primary energy consumption (0.971) results in higher CO2 emission (equations (c) and (f)). A one-percent increase in the CO2 emissions leads to a one-percent       

10 As shown in Table 2, lnOC and lnPEC are endogenous. Hence, both variables are dropped. The F-statistic of ARDL bounds test (with 3 lags) is 15.591, which rejects the null of no cointegration.

(1.1 and 1.0) increase in energy consumption of oil (lnOC) and primary electricity (lnPEC), respectively (equations (b) and (e)). This study caters the endogeneity in the independent variables (Table 2), the estimates of fully modified ordinary least squares (FMOLS) and dynamic ordinary least squares (DOLS) are computed and reported in Appendix B. In general, their estimates are robust to the ARDL estimator, especially the FMOLS.

Table 4 presents the estimated ECM of ARDL equations (a), (c) and (d) as suggested by augmented production function (i.e. lnY –lnOC -lnCO2, and lnY- lnPEC -lnCO2), and CO2 emission equation (i.e. lnCO2 –lnOC –lnY). The remaining ARDL specifications (b), (e) and (f) are not covered since the dynamic lag structure is ARDL (0,0,0) indicating no short-run shocks. Both equations (a’) and (d’) suggest that the one and two years lagged CO2 emissions growth lead to reduce the Cambodian economic growth, in the short-run. Their estimated short- run elasticities are ranged between -0.199 and -0.391. The CO2 equation shows, a 1% of OC increases will lead 0.85% additional CO2 emissions to Cambodia. Again, the estimated error correction terms (ECTt-1) which measure the speed of adjustment to equilibrium, are significant and in an expected sign. It further confirms the existence of a long-run relationship among real GDP, energy consumption, and CO2 emissions. The estimated value of -0.093 (or -0.091) suggests a speed of convergence to equilibrium of about 9% per year or approximately 11 years to equilibrium. The ECT of CO2 equation (c’) is statistically insignificant further supporting the finding of no cointegration as obtained from the bound test (Table 2). Figure 2 presents the CUSUM and CUSUMQ plots for ECM equations. The CUSUM tests suggest stability (within 5% critical bounds), while the CUSUMQ of equations (c’) and (d’) reveal unstable the estimated coefficients.

The Toda and Yamamoto’s (1995) testing method is used because the underlying variables are non-stationary, or mixture in their degree of integration, I(d).¹¹ Table 5 presents the empirical results, which takes into account the statement of ‘the cause occurs before the effect…’ by Granger (1988). In panel I, the test statistic, 9.758 (or 12.61) does reject the null hypothesis of lnY does not Granger cause lnOC (or lnCO2) at 5%. Also, the null hypothesis of lnCO2 does not Granger cause lnY is rejected (see the last column). The remaining test statistics are statistically insignificant. It shows that the causality is unidirectional which runs from real GDP to oil consumption. There is no reversed causality. This finding is in line with the standard consumption function that relates energy consumption to income (real GDP). A bidirectional causality is confirmed between real GDP and CO2 emissions. It supports the EKC hypothesis that relates the pollution (CO2) to output. Similar findings are obtained on primary energy consumption, lnPEC as the test statistics reported in panel II. The real GDP does       

11 This method has been widely employed by researchers. Given the finite annual observations in this study, the critical values of the causality tests are obtained by using the bootstrap modified Wald statistics critical values by Hacker and Hatemi-J (2006). Therefore, the reported results do not suffer the finite sample issue. The VECM can be applied for robustness check if the assumption that all variables are non-stationary, or I(1).

Granger cause the primary energy consumption. A bidirectional causality is obtained between real GDP and CO2 emissions. In panel III, no causality between the electricity net consumption (lnENC) and real GDP. Again, bidirectional causality is obtained between real GDP and CO2 which is consistent with panels I and II. Real GDP is an ‘intermediator’ from CO2 emissions to energy consumption. In general, the entire empirical results are diagrammatically in Figure 3.

Table 2: ARDL bound F-test for cointegration.

F-statistics (a) F(lnY| lnOC, lnCO2) 18.417^***

(b) F(lnOC|lnY, lnCO2) 4.496^**

(c) F(lnCO2| lnOC, lnY) 3.252

(d) F(lnY|lnPEC, lnCO2) 17.982^***

(e) F(lnPEC|lnY, lnCO2) 4.826^**

(f) F(lnCO2| lnPEC, lnY) 3.397

Narayan’s (2005) critical values bound of the F-statistics:

Intercept and no trend (T=30, k=2) Lower bound, I(0) Upper bound, I(1)

90% 2.915 3.695

95% 3.538 4.428

99% 5.155 6.265

Notes: Asymptotic critical bound are obtained from Narayan (2005). Here, the ‘k’ is the number of regressors. The (^***) and (^**) denote significance levels at 1% and 5%, respectively. Given a sample of 31 observations (1980-2010), a lag length of 3 is implemented or ARDL (3,3,3). Enders (2014) proposed a lag length that is maximally T^1/3 where T is the number of observations.

Table 3: ARDL long-run elasticities.

ARDL (a) lnY| lnOC,

lnCO2

(b) lnOC|lnY,

lnCO2

(c) lnCO2| lnOC, lnY

(d) lnY|lnPEC,

lnCO2

(e) lnPEC|lnY,

lnCO2

(f) lnCO2| lnPEC, lnY Intercept 1.910

(0.974)

-7.292^***

(0.000)

6.488^***

(0.000)

-252.3 (0.263)

-11.46^***

(0.000)

11.145^***

(0.000)

lnCO2 3.878

(0.651)

1.105^***

(0.000) - 24.71 (0.226)

1.028^***

(0.000) -

lnY - -0.009

(0.248)

-0.009

(0.907) - -0.001 (0.830)

0.001 (0.714)

lnOC -4.476

(0.572) - 0.970^***

(0.001) - - -

lnPEC - - - -25.11

(0.216) - 0.971^***

(0.000) Diagnostic Chi-squared test statistics - Lagrange Multiplier (LM) Version

Serial correlation 1.151 4.229^** 0.127 3.348^* 0.261 0.256 Functional form 6.737^*** 3.601^* 0.000 5.065^** 2.019 2.666^*

Table 3 (continued).

ARDL (a) lnY| lnOC,

lnCO2

(b) lnOC|lnY,

lnCO2

(c) lnCO2| lnOC, lnY

(d) lnY|lnPEC,

lnCO2

(e) lnPEC|lnY,

lnCO2

(f) lnCO2| lnPEC, lnY Diagnostic Chi-squared test statistics - Lagrange Multiplier (LM) Version

Normality 1.111 163.13^*** 96.97^*** 1.980 0.300 0.295 Heteroscedasticity 0.073 2.004 1.716 0.039 3.067^* 2.715^*

Notes: lnY|lnOC, lnCO2 is interpreted as lnY being influenced by lnOC and lnCO2 and the same applies to the rest of the equations. The estimated coefficients are reported with the p-values in parenthesis. (^***) (^**) and (^*) denote significance levels at 1%, 5% and 10%, respectively. The sample periods are 1984-2010 for equations (a), (b) and (d), and 1985-2010 for equations (c), (e) and (f) after initial ARDL (3,3,3) computed by Microfit.

Table 4: Error correction model (ECM) for the selected ARDL model.

Independent variables

(a’)

ARDL(lnY|lnOC, lnCO2) ARDL (1,0,3)

(c’)

ARDL(lnCO2|lnOC, lnY) ARDL (1,1,0)

(d’)

ARDL(lnY|lnPEC, lnCO2) ARDL (1,0,3) Intercept 0.177

(0.973) 0.902

(0.749)

-23.089 (0.234)

∆lnCO2 0.039

(0.963)

- 1.875 (0.272)

∆lnCO2t-1 -0.217^**

(0.030) - -0.199^**

(0.037)

∆lnCO2t-2 -0.391^***

(0.000) - -0.383^***

(0.000)

∆lnOC -0.417

(0.569) 0.850^***

(0.000)

-

∆lnPEC - - -2.297

(0.177)

∆lnY - -0.001 (0.876)

-

ECTt-1 -0.093^***

(0.004) -0.139

(0.743)

-0.091^***

(0.004) Notes: The estimated coefficients are reported with the p-values in parenthesis. (^***) and (^**) denote significance levels at 1% and 5%, respectively. The lag structures of the ARDL (.) equation are selected by Schwarz Bayesian criteria (SBC) for the short-run dynamic. The first-differenced variable (denoted as ∆) is the difference between current value and lagged one value, e.g., ∆lnCO2

= lnCO2t – lnCO2t-1. The ECT for equations (a), (c) and (d) are ecm_1 = lnY + 4.476*lnOC - 3.878*lnCO2 - 1.91, ecm_3 = lnCO2 -0.970*lnOC + 0.009*lnY - 6.488, and ecm_4 = lnY + 25.106*lnPEC - 24.709*lnCO2 + 252.336, respectively.

Figure 2: CUSUM and CUSUMQ tests.

Table 5: Toda and Yamamoto (1995) non-causality test with the bootstrap approach.

Independent variables

Panel I lnY lnOC lnCO2

lnY - 1.970 9.795^**

lnOC 9.758^** - 2.000

lnCO2 12.610^** 2.127 - Panel II lnY lnPEC lnCO2

lnY - 1.700 9.470^**

lnPEC 10.666^** - 2.839

lnCO2 10.685^** 1.183 -

CUSUMQ for eq. 4

The straight lines represent critical bounds at 5% significance level -0.5

0.0 0.5 1.0 1.5

1983 1988 1993 1998 2003 2008

CUSUM for eq. 3

The straight lines represent critical bounds at 5% significance level -5

-10 -15 0 5 10 15

1985 1990 1995 2000 2005 2010

CUSUMQ for eq. 3

The straight lines represent critical bounds at 5% significance level -0.5

0.0 0.5 1.0 1.5

1985 1990 1995 2000 2005 2010

CUSUM for eq. 4

The straight lines represent critical bounds at 5% significance level -5 

-10 

-15 

0 5 10 15

1983 1988 1993  1998 2003 2008

CUSUM for eq. 1

The straight lines represent critical bounds at 5% significance level -5

-10 -15 0 5 10 15

1983 1988 1993 1998 2003 2008

CUSUMQ for eq. 1

The straight lines represent critical bounds at 5% significance level -0.5

0.0 0.5 1.0 1.5

1983 1988 1993 1998 2003 2008

Table 5 (continued).

Independent variables

Panel III lnY lnENC lnCO2

lnY - 0.322 6.287^*

lnENC 2.472 - 1.882

lnCO2 6.958* 0.183 -

Notes: The reported value are the Wald statistics. (^**) and (^*) indicate 5%, and 10% significance levels based on the bootstrap modified Wald statistic (Hacker & Hatemi-J, 2006). The critical values obtained for 1%, 5% and 10% significance levels are 17.888, 10.4992 and 7.892, respectively. If the variables are significant, the column variable Granger causes the row variable.

The selected lag, k are based on the SBC values suggested. Panels I and II follows the VAR(k + dmax) structure of VAR(2 +1). For panel III, the VAR(lnY= 1 +1, lnCO2 = 1 +1, lnENC = 1 +2) since the lnENC is I(2).

Notes: The solid line represents the Granger non-causality results from the Toda-Yamamoto method. The bold dash line represents the effects in the long-run; the small dash line represents the short-run effect.

Figure 3: Summary of inter-linkages among energy consumption, real GDP, and CO2 emissions.

Oil Consumption

Electricity Net Consumption

Primary Energy Consumption

Real GDP

 

CO2 Emissions 0.970 1.105

1.03

0.971 -0.391 to -0.199

0.850

5. Concluding remarks

This study contributes to the existing literature by delivering fresh evidence of the energy-growth-CO2 nexus for a transition economy of MRC countries, Cambodia. This study finds that:

a) There is, at least a long-run relationship (cointegration) among GDP, energy consumption (oil consumption and primary energy consumption), and CO2

emissions. Oil and primary energy consumptions respectively increase CO2

emissions. Both energy consumption and CO2 emissions have no impact on the Cambodian GDP in the long-run.

b) The CO2 variable has negative short-run implication on GDP, while oil consumption results in additional CO2 emissions. The speed of adjustment is approximately 11 years in order to achieve a long-run equilibrium among the variables.

c) GDP does Granger cause energy consumption. A bidirectional causality between GDP and CO2 emissions. The identified transmission channel for CO2 emissions to energy consumption (oil and primary) is through GDP.

These findings are relevant for policy implication. As projected that Cambodia’s energy consumption is growing at an average of 5.2% per year between 2009 and 2035 that the Cambodian energy consumption can be reduced to 4.3% with an overall reduction of future energy demand of 20% by 2035.¹² Hence, energy policies to cut the consumption of primary and oil energy can be implemented in order to lower the CO2 emissions, in the long-run 3 million tonnes of CO2 in 2035. The time cost of energy and CO2 mismanagement is 11 years (i.e., the speed of adjustment) to the Cambodian government to allow the national energy efficiency policy in results.

From the non-causality finding, energy policies of reducing either oil or primary energy consumption can be implemented without deteriorating the country’s output. The Cambodian demand for energy is caused by GDP that the recent high growth (approximately 7%) scenario requires more energy inputs to support the core growth sectors. In this context, a wider understanding of these sectors is needed on their CO2 emissions and their strategy towards green energy.

The national energy efficiency policy is currently under development. Therefore, reduction in oil and primary energy consumption can be achieved by substitution of renewable energy or clean energy. Government and private sectors are suggested to employ advanced technology - carbon-free power, and renewable energy for achieving environmental friendly and promoting economic development in the future. The tax credit can be implemented for those industries using renewable energy. According to Sarraf et al. (2013, p. 228), Cambodia is categorised as one of the richest economies with natural energy resources such as solar, wind, biomass, and hydropower among the developing countries. The Cambodian renewable energy resources can generate up to 67,388 GWh energy

      

12 See http://www.phnompenhpost.com/business/move-lift-energy-efficiency-cambodia (Accessed July 2, 2014).

per year. They also found that the best option for rural electrification is renewable energy resources.

A few of concerns are necessary for further study in the field. Ozturk (2010) suggested that new approaches and perspectives are important for further study rather than by simply applying traditional econometric methods, adding new variables, using different countries, and different time intervals.

References

Abbasi, F., & Riaz, K. (2016). CO2 emissions and financial development in an emerging economy: An augmented VAR approach. Energy Policy, 90(March), 102–114.

Abdul, J. (2014). Energy–growth conundrum in energy exporting and importing countries: Evidence from heterogeneous panel methods robust to cross- sectional dependence. Energy Economics, 44(July), 314–324.

Ahmed, K., Shahbaz, M., & Kyophilawong, P. (2016). Revisiting the emissions- energy-trade nexus: Evidence from the newly industrializing countries.

Environmental Science and Pollution Research, (In Press), 1-16.

doi:10.1007/s11356-015-6018-x.

Ajmi, A. N., Hammoudeh, S., Nguyen, D. K., & Sato, J. R. (2015). On the relationships between CO2 emissions, energy consumption and income:

The importance of time variation. Energy Economics, 49(May), 629–638.

Alam, A. (2013). Nuclear energy, CO2 emissions and economic growth: The case of developing and developed countries. Journal of Economic Studies, 40(6), 822-834.

Al-Mulali, U. (2015). The impact of biofuel energy consumption on GDP growth, CO2 emission, agricultural crop prices, and agricultural production.

International Journal of Green Energy, 12(11), 1100-1106.

Al-Mulali, U., & Tang, C. F. (2013). Investigating the validity of pollution haven hypothesis in the gulf cooperation council (GCC) countries. Energy Policy, 60(September), 813-819.

Al-Mulali, U., Choong, W.-W., & Low, S.-T. (2015a). Investigating the environmental Kuznets curve (EKC) hypothesis by utilizing the ecological footprint as an indicator of environmental degradation. 48(January), 315–

323.

Al-Mulali, U., Ozturk, I., & Hooi, L. H. (2015c). The influence of economic growth, urbanization, trade openness, financial development, and renewable energy on pollution in Europe. Natural Hazards, 79(1), 621-644.

Al-Mulali, U., Saboori, B., & Ozturk, I. (2015b). Investigating the environmental Kuznets curve hypothesis in Vietnam. Energy Policy, 73(January), 123–

131.

 

Al-Mulali, U., Solarin, S. A., & Ozturk, I. (2016). Investigating the presence of the environmental Kuznets curve (EKC) hypothesis in Kenya: An autoregressive distributed lag (ARDL) approach. Natural Hazards, 80(3), 1729-1747.

Alshehry, A. S., & Belloumi, M. (2015). Energy consumption, carbon dioxide emissions and economic growth: The case of Saudi Arabia. Renewable and Sustainable Energy Reviews, 41(January), 237–247.

Ang, J. B. (2007). CO2 emissions, energy consumption, and output in France.

Energy Policy, 35(10), 4772–4778.

Apergis, N. (2016). Environmental Kuznets curves: New evidence on both panel and country-level CO2 emissions. Energy Economics, 54(February), 263- 271.

Araç, A., & Mübariz, H. (2014). Asymmetries in the dynamic interrelationship between energy consumption and economic growth: Evidence from Turkey. Energy Economics, 44(July), 259–269.

Asafu-Adjaye, J. (2000). The relationship between energy consumption, energy prices and economic growth: Time series evidence from Asian developing countries. Energy Economics, 22(6), 615-625.

Asongu, S., Montasser, G. E., & Toumi, H. (2015). Testing the relationships between energy consumption, CO2 emissions, and economic growth in 24 African countries: A panel ARDL approach. Environmental Science and Pollution Research, (In Press), 1-11. doi:10.1007/s11356-015-5883-7.

Bae, j. H., Li, D. D., & Rishi, M. (2016). Determinants of CO2 emission for post- Soviet Union independent countries. Climate Policy, In Press, 1-26.

doi:10.1080/14693062.2015.1124751.

Baek, J. (2016). A new look at the FDI-income-energy-environment nexus:

Dynamic panel data analysis of ASEAN. Energy Policy, 91(April), 22-27.

Balaguer, J., & Cantavella, M. (2016). Estimating the environmental Kuznets curve for Spain by considering fuel oil prices (1874-2011). Ecological Indicators, 60(January), 853-859.

Bastola, U., & Sapkota, P. (2015). Relationships among energy consumption, pollution emission, and economic growth in Nepal. Energy, 80(February), 254–262.

Begum, R. A., Sohag, K., Syed Abdullah, S. M., & Jaafar, M. (2015). CO2

emissions, energy consumption, economic and population growth in Malaysia. Renewable and Sustainable Energy Reviews, 41(January), 594–601.

Bento, J. P., & Moutinho, V. (2016). CO2 emissions, non-renewable and renewable electricity production, economic growth, and international trade in Italy. Renewable and Sustainable Energy Reviews, 55(March), 142-155.

Burke, P. J., Shahiduzzaman, M., & Stern, D. I. (2015). Carbon dioxide emissions in the short run: The rate and sources of economic growth matter. Global Environmental Change, 33(July), 109–121.

Chandran, V., & Tang, C. F. (2013a). The dynamic links between CO2 emissions, economic growth and coal consumption in China and India. Applied Energy, 104(April), 310-318.

Chandran, V., & Tang, C. F. (2013b). The impacts of transport energy consumption, foreign direct investment and income on CO2 emissions in ASEAN-5 economies. Renewable and Sustainable Energy Reviews, 24(August), 445-453.

Dickey, D. F., & Fuller, W. A. (1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of American Statistical Association, 74(366), 427-431.

Dogan, E., & Seker, F. (2016). The influence of real output, renewable and non- renewable energy, trade and financial development on carbon emissions in the top renewable energy countries. Renewable and Sustainable Energy Reviews, 60(July), 1074-1085.

Dogan, E., & Turkekul, B. (2016). CO2 emissions, real output, energy consumption, trade, urbanization and financial development: Testing the EKC hypothesis for the USA. Environmental Science and Pollution Research, 23(2), 1203-1213.

Enders, W. (2014). Applied Econometric Time Series (4nd Ed.). New York: Wiley.

Engle, R. F., & Granger, C. W. (1987). Cointegration and error correction:

Representation, estimation, and testing. Econometrica, 55(2), 251-276.

Farhani, S., & Ozturk, I. (2015). Causal relationship between CO2 emissions, real GDP, energy consumption, financial development, trade openness, and urbanization in Tunisia. Environmental Science and Pollution Research, 22(20), 15663-15676.

Fujii, H., & Managi, S. (2016). Economic development and multiple air pollutant emissions from the industrial sector. Environmental Science and Pollution Research, 23(3), 2802-2812.

Georgiev, E., & Mihaylov, E. (2015). Economic growth and the environment:

Reassessing the environmental Kuznets curve for air pollution emissions in OECD countries. Letters in Spatial and Resource Sciences, 8(1), 29-47.

Granger, C. W. (1988). Some recent developments in a concept of causality.

Journal of Econometrics, 39(1-2), 199-211.

Hacker, R. S., & Hatemi-J, A. (2006). Tests for causality between integrated variables using asymptotic and bootstrap distributions: Theory and application. Applied Economics, 38(13), 1489-1500.

Haldrup, N. (1998). An econometric analysis of I(2) variables. Journal of Economic Surveys, 12(5), 595–650.

Hao, Y., & Liu, Y.-M. (2015). Has the development of FDI and foreign trade contributed to China's CO2 emissions? An empirical study with provincial panel data. Natural Hazards, 76(2), 1079-1091.

Haseeb, M., & Azam, M. (2015). Energy consumption, economic growth and CO2

emission nexus in Pakistan. Asian Journal of Applied Sciences, 8(1), 27- 36.

Heidari, H., Katircioglu, S. T., & Saeidpour, L. (2015). Economic growth, CO2

emissions, and energy consumption in the five ASEAN countries.

International Journal of Electrical Power and Energy Systems, 64(January), 785–791.

Hornok, A., & Larsson, R. (2000). The finite sample distribution of the KPSS test.

Econometrics Journal, 3(1), 108-121.

Jammazi, R., & Aloui, C. (2015a). On the interplay between energy consumption, economic growth and CO2 emission nexus in the GCC countries: A comparative analysis through wavelet approaches. Renewable and Sustainable Energy Reviews, 51(November), 1737–1751.

Jammazi, R., & Aloui, C. (2015b). Environment degradation, economic growth and energy consumption nexus: A wavelet-windowed cross correlation approach. Physica A, 436(October), 110–125.

Javid, M., & Sharif, F. (2016). Environmental Kuznets curve and financial development in Pakistan. Renewable and Sustainable Energy Reviews, 54(February), 406-414.

Jebli, M. B., & Youssef, S. B. (2015). The environmental Kuznets curve, economic growth, renewable and non-renewable energy, and trade in Tunisia.

Renewable and Sustainable Energy Reviews, 47(July), 173–185.

Jebli, M. B., Youssef, S. B., & Ozturk, I. (2016). Testing environmental Kuznets curve hypothesis: The role of renewable and non-renewable energy consumption and trade in OECD countries. Ecological Indicators, 60(January), 824-831.

Jiranyakul, K. (2016). Causal linkages between electricity consumption and GDP in Thailand: Evidence from the bounds test. Economics Bulletin, 36(2), A90.

Johansen, S., & Juselius, K. (1990). Maximum likelihood estimation and inference on cointegration - with applications to the demand for money.

Oxford Bulletin of Economics and Statistics, 52(2), 169-210.

Kais, S., & Sami, H. (2016). An econometric study of the impact of economic growth and energy use on carbon emissions: Panel data evidence from fifty eight countries. Renewable and Sustainable Energy Reviews, 59(June), 1101-1110.

Kasman, A., & Duman, Y. S. (2015). CO2 emissions, economic growth, energy consumption, trade and urbanization in new EU member and candidate countries: A panel data analysis. Economic Modelling, 44(January), 97–

103.

Katrakilidis, C., Kyritsis, I., & Patsika, V. (2016). The dynamic linkages between economic growth, environmental quality and health in Greece. Applied Economics Letters, 23(3), 217-221.

Khan, M. M., Zaman, K., Irfan, D., Awan, U., Ali, G., Kyophilavong, P., Shahbaz, M., & Naseem, I. (2016). Triangular relationship among energy consumption, air pollution and water resources in Pakistan. Journal of Cleaner Production, 112(Part 2 (January)), 1375-1385.

Kraft, J., & Kraft, A. (1978). On the relationship between energy and GNP.

Journal of Energy Development, 3(2), 401-403.

Kwiatkowski, D., Phillips, P. C., Schmidt, P., & Shin, Y. (1992). Testing the null hypothesis of stationarity against the alternative of a unit root. Journal of Econometrics, 54(1-3), 159-178.

Lee, J. W., & Brahmasrene, T. (2014). ICT, CO2 emissions and economic growth:

Evidence from a panel of ASEAN. Global Economic Review, 43(2), 93-109.

Lee, S., & Chong, W. O. (2016). Causal relationships of energy consumption, price, and CO2 emissions in the U.S. building sector. Resources, Conservation and Recycling, 107(February), 220-226.

Li, T., Wang, Y., & Zhao, D. (2016). Environmental Kuznets curve in China: New evidence from dynamic panel analysis. Energy Policy, 91(April), 138-147.

Liew, V. K.-S., Thurai, M. N., & Wong, W.-K. (2012). Are sectoral outputs in Pakistan led by energy consumption? Economics Bulletin, 32(3), 2326-31.

Long, X., Naminse, E. Y., Du, J., & Zhuang, J. (2015). Nonrenewable energy, renewable energy, carbon dioxide emissions and economic growth in China from 1950 to 2012. Renewable and Sustainable Energy Reviews, 52(December), 680–688.

Mohammadi, H., & Parvaresh, S. (2014). Energy consumption and output:

Evidence from a panel of 14 oil-exporting countries. Energy Economics, 41(January), 41-46.

Narayan, P. K. (2005). The saving and investment nexus for China: Evidence from cointegration tests. Applied Economics, 37(17), 1979-1990.

Omri, A., Daly, S., Rault, C., & Chaibi, A. (2015). Financial development, environmental quality, trade and economic growth: What causes what in MENA countries. Energy Economics, 48(March), 242–252.

Onafowora, O. A., & Oluwole, O. (2014). Bounds testing approach to analysis of the environment Kuznets curve hypothesis. Energy Economics, 44(July), 47-62.

Ozturk, I. (2010). A literature survey on energy–growth nexus. Energy Policy, 38(1), 340-349.

Pesaran, H. M., & Pesaran, B. (1997). Working with Microfit 4.0 Interactive Econometric Analysis. New York: Oxford University Press Inc.

Pesaran, M., Shin, Y., & Smith, R. (2001). Bounds testing approaches to the analysis of level relationships. Journal of Applied Econometrics, 16(3), 289-326.

Phillips, P. C., & Perron, P. (1988). Testing for a unit root in time series regression. Biometrika, 75(2), 335-346.

Robalino-Lopez, A. R., Nieto, A. M., & Ramos, J. E. (2015). Studying the relationship between economic growth, CO2 emissions, and the environmental Kuznets curve in Venezuela (1980-2025). Cover image Renewable and Sustainable Energy Reviews, 41(January), 602–614.

Royal Government of Cambodia. (2013). National policy, strategy and action plan on energy efficiency in Cambodia. Phnom Penh: Ministry of Industry, Mines, and Energy. Retrieved 5 7, 2014, from http://www.euei- pdf.org/sites/default/files/files/field_pblctn_file/EUEI%20PDF_Cambo dia_Energy%20Efficiency_May2013_EN.pdf.

Ruhul, A. S., Kamrul, H., & Sahar, S. (2014). Renewable and non-renewable energy consumption and economic activities: Further evidence from OECD countries. Energy Economics, 44(July), 350–360.

Saidi, K., & Hammami, S. (2015). The impact of energy consumption and CO2

emissions on economic growth: Fresh evidence from dynamic simultaneous-equations models. Sustainable Cities and Society, 14(February), 178–186.

Saidi, K., & Hammami, S. (2016). Economic growth, energy consumption and carbone dioxide emissions: Recent evidence from panel data analysis for 58 countries. Quality and Quantity, 50(1), 361-383.

Saidi, K., & Mbarek, M. B. (2016). The impact of income, trade, urbanization, and financial development on CO2 emissions in 19 emerging economies.

Environmental Science and Pollution Research, 1-10.

doi:10.1007/s11356-016-6303-3.

Salahuddin, M., Gow, J., & Ozturk, I. (2015). Is the long-run relationship between economic growth, electricity consumption, carbon dioxide emissions and financial development in Gulf Cooperation Council Countries robust?

Renewable and Sustainable Energy Reviews, 51(November), 317–326.

Sarraf, M., Rismanchi, B., Saidur, R., Ping, H., & Rahim, N. (2013). Renewable energy policies for sustainable development in Cambodia. Renewable and Sustainable Energy Reviews, 22(June), 223–229.

Sumabat, A. K., Lopez, N. S., Yu, K. D., Hao, H., Li, R., Geng, Y., & Chiu, A. S.

(2016). Decomposition analysis of Philippine CO2 emissions from fuel combustion and electricity generation. Applied Energy, 164(February), 795-804.

Tang, C. F. (2008). A re-examination of the relationship between electricity consumption and economic growth in Malaysia. Energy Policy, 36(8), 3077-3085.

Tang, C. F. (2009). Electricity consumption, income, foreign direct investment, and population in Malaysia: New evidence from multivariate framework analysis. Journal of Economic Studies, 36(4), 371-382.

Tang, C. F., & Muhammad, S. (2013). Sectoral analysis of the causal relationship between electricity consumption and real output in Pakistan. Energy Policy, 60(September), 885-891.

Tang, C. F., & Salah, A. (2014). The impacts of tourism, energy consumption and political instability on economic growth in the MENA countries. Energy Policy, 68, 458-464.

Tang, C. F., & Tan, B. W. (2014). The linkages among energy consumption, economic growth, relative price, foreign direct investment, and financial development in Malaysia. Quality and Quantity, 48(2), 781-797.

Tang, C. F., & Tan, B. W. (2015). The impact of energy consumption, income and foreign direct investment on carbon dioxide emissions in Vietnam.

Energy, 79(January), 447–454.

Tang, C. F., & Tan, E. C. (2012). Electricity consumption and economic growth in Portugal: Evidence from a multivariate framework analysis. Energy Journal, 33(4), 23-48.

Tang, C. F., & Tan, E. C. (2013). Exploring the nexus of electricity consumption, economic growth, energy prices and technology innovation in Malaysia.

Applied Energy, 104(April), 297-305.

Tang, C. F., Muhammad, S., & Mohamed, A. (2013). Re-investigating the electricity consumption and economic growth nexus in Portugal. Energy Policy, 62, 1515-1524.

Tang, T. C., & Chea, R. (2013). Export-led growth in Cambodia: An empirical study. Economics Bulletin., 33(1), 655-662.

Toda, H. Y., & Yamamoto, T. (1995). Statistical inference in vector autoregressions with possibly integrated processes. Journal of Econometrics, 66(1-2), 225-250.

United Nations. (2003). An investment guide to Cambodia: Opportunities and conditions. New York and Geneva: United Nations. Retrieved 5 6, 2014, from http://unctad.org/en/Docs/iteiia20036_en.pdf.

Wang, Y., Han, R., & Kubota, J. (2016b). Is there an environmental Kuznets curve for SO2 emissions? A semi-parametric panel data analysis for China.

Renewable and Sustainable Energy Reviews, 54(February), 1182-1188.

Wang, Y., Lili, C., & Kubota, J. (2016a). The relationship between urbanization, energy use and carbon emissions: Evidence from a panel of Association of Southeast Asian Nations (ASEAN) countries. Journal of Cleaner Production, 112(January), 1368-1374.

Wolde-Rufael, Y. (2014). Electricity consumption and economic growth in transition countries: A revisit using bootstrap panel Granger causality analysis. Energy Economics, 44(July), 325-330.

Yorucu, V. (2016). Growth impact of CO2 emissions caused by tourist arrivals in Turkey. International Journal of Climate Change Strategies and Management, 8(1), 19-37.

Yu, B., Li, X., Qiao, Y., & Shi, L. (2015). Low-carbon transition of iron and steel industry in China: Carbon intensity, economic growth and policy intervention. Journal of Environmental Sciences, 28(February), 137–147.

Yuan, X., Mu, R., Zuo, j., & Wang, Q. (2015). Economic development, energy consumption, and air pollution: A critical assessment in China. Human and Ecological Risk Assessment: An International Journal, 21(3), 781- 798.

Zaman, K., Abdullah, A., Khan, A., Moh Nasir, M. R., Tengku Hamzah, T. A., &

Hussain, S. (2016a). Dynamic linkages among energy consumption, environment, health and wealth in BRICS countries: Green growth key to sustainable development. Renewable and Sustainable Energy Reviews, 56(April).

Zaman, K., Shahbaz, M., Loganathan, N., & Raza, S. A. (2016b). Tourism development, energy consumption and environmental Kuznets curve:

Trivariate analysis in the panel of developed and developing countries.

Tourism Management, 54(June), 275-283.

Zhang, L., & Gao, j. (2016). Exploring the effects of international tourism on China's economic growth, energy consumption and environmental pollution: Evidence from a regional panel analysis. Renewable and Sustainable Energy Reviews, 53(January), 225-234.

Zhang, Y.-J., & Da, Y.-B. (2015). The decomposition of energy-related carbon emission and its decoupling with economic growth in China. Renewable and Sustainable Energy Reviews, 41(January), 1255–1266.

Ziaei, S. M. (2015). Effects of financial development indicators on energy consumption and CO2 emission of European, East Asian and Oceania countries. Renewable and Sustainable Energy Reviews, 42(February), 752–759.